Although long term (24-hour) electrocardiographic monitoring of ambulatory patients has become an integral part of medical diagnosis, analysis of data collected has advanced little since introduction of the screening procedure by Holter in 1957. Although digital processing using traditional pattern recognition techniques and interval analysis has been proposed by several groups of investigators, we are not aware of a sufficiently reliable computerized system of Holter analysis for use in clinical practice. A newer, more general procedure, syntactical pattern analysis, will be employed in the proposed work. This technique permits machine coding of elemental electrocardiographic segments (primitives) of each PQRST complex as well as the intervals between beats. Arrhythmias will then be described in terms of these primitives in accord with the rules (grammars) governing colligative properties of the primitives. A statistical approach to the syntactical method has been developed to accomodate beat-to-beat variability in the ECG morphology and rhythm. The proposal will utilize our library of approximately 300 physician interpreted Holter tapes. The design and construction of a pre-processor is necessary to enable both the continuous processing of 24-hour Holter tapes at 120 times real time and the use of available, but limited, computer facilities.